You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For most APIs we can just use Vector2 still (e.g. circle position, Box2, etc) but transform should be using a position struct that can be swapped out between 32-64bit precision (as to how we deliver it, separate engine builds or smth IDK but I think the performance difference is immaterial enough to want double-precision).
For most APIs we can just use Vector2 still (e.g. circle position, Box2, etc) but transform should be using a position struct that can be swapped out between 32-64bit precision (as to how we deliver it, separate engine builds or smth IDK but I think the performance difference is immaterial enough to want double-precision).
See https://github.com/erincatto/box2d/blob/56edae79f2949d86142b03450d5d60f63bcf5a6f/docs/large_worlds.md